Showing posts with label ModelMuse. Show all posts
Showing posts with label ModelMuse. Show all posts

Thursday, January 21, 2021

Дайджест ссылок

1st Groundwater Project Event. Скоро должна начаться первая онлайновая конференция от Groundwater Project. Спешите записаться.
The first ever Groundwater Project Event will be 100% online between the 3rd and 26th of February 2021. A series of meetings and discussions will be promoted, addressing relevant and innovative themes presented by the best references in hydrogeology in the world. The event will be broadcasted on the Groundwater Project YouTube channel and the public will be able to interact with the speakers through the live chat. All presentations will be broadcasted in English or with English subtitles. Курсы по ГИС. В основном речь идет о qGIS.

A comprehensive list of specific Python packages for hydrogeology and groundwater modeling. Список утилит на Python-е для решения прикладных задач в гидрогеологии и геофильтрационного моделирования.
Development of open source software brings amazing new tools in all fields. In hydrogeology and groundwater modeling there is an increasing number of specific open source software and programming packages. We wanted to compile the latest libraries for Python related to hydrogeology, we asked for references and researched over the web to provide you the following list.
Lecture Notes. Лекции по гидрогеологии, геогидродинамике и геомиграции.
Some of the figures in these lecture notes are adapted from or inspired by illustrations in Dingman, S. Lawrence. Physical Hydrology. 2nd edition. Prentice-Hall, 2002.
Estimation of optimal complexity in groundwater models using cross-validation methods. Очень любопытная статья, в которой в частности идет речь об очень забавном феномене, который я сам неоднократно наблюдал на своем опыте (но никогда не пытался подвести под него теоретическую базу — да и куда мне, я больше практик): по мере увеличения сложности и «комплексности» модели, точность прогнозов, сделанных на её основе, сначала закономерно повышается, а потом начинает почему-то снижаться. Там приведен очень наглядный пример со степенными и линейными аппроксимациями в экселе.
The level of complexity that a groundwater model should have is an important and recurrent question for hydrogeologists. A few weeks ago an interesting discussion about this topic occurred on LinkedIn, with one of the main conclusions being that the optimal complexity should minimize predictive error/uncertainty by improving the fit to the existing observations, without overfitting them, as described in Figure 1. Understanding the meaning and implications of this figure is key for the modelling process. In that sense, I decided to implement two sets of models to illustrate how different degrees of complexity affect predictive error. We will also see how to estimate the optimal level of complexity in groundwater models utilizing cross-validation techniques commonly used by data scientists.
ModelMuse Videos. Видеоуроки по работе с ModelMuse (бесплатный пост-препроцессор для MODFLOW, если кто не знал). Программа очень мощная, но learning curve у нее просто запредельная.